SPH Modelling of the 3D laboratory experiments on the Vajont landslide

The Vajont disaster occurred on October 9, 1963: a huge landslide fell into the Erto hydroelectric reservoir and generated a wave which overtopped the Vajont dam, sweeping away the downstream village of Longarone and causing about 2000 casualties. Before the catastrophic event, some experiments were performed at the Research Centre of Nove (Treviso, Italy) on a 3D physical model of the Vajont reservoir. Owing to the uncertainties on the complex physical mechanism (i.e. volume of the landslide and its kinematic), the energy of the generated wave was underestimated. These experimental data, which were made available recently, have been considered to perform numerical experiments of the Vajont landslide with the Smoothed Particle Hydrodynamics (SPH). SPH is a well-established Lagrangian mesh-free particle method initially developed for astrophysical applications (Lucy, 1977; Gingold & Monaghan, 1977) and subsequently successfully extended to free-surface flows (Monaghan 1994). SPH proved to be effective in a number of different applications, e.g. for simulating coupled soil-water dynamics in a flushing manoeuvre (Manenti et al., 2012a) and for analyzing the 3D dynamics of a tsunami wave propagating on a beach (Guandalini et al., 2013). The study has been carried out using the code SPHERA (Agate et al., 2010), developed at RSE in the frame of the SPH project funded by Energy Research System. The main code features include the continuum discretization by means of a finite number of material particles carrying physical properties and moving according to Newton’s equations of the classical physics, as the traditional SPH. Spatial derivatives of a variable at a point are approximated by using the information on the neighboring particles based on the kernel approximation. Other basics of the code concern innovative developments of many modelling aspects as the boundary treatment (Di Monaco et al., 2011), the time integration scheme and the interactions among particles with different material properties (Manenti et al., 2012b). The geometrical complexity of the 3D model provided to SPHERA, required the development of a proper pre-processing methodology including digitalization and rendering of the terrain maps before and after the disaster, the generation of corresponding contour lined surfaces and the generation of final 3D geometry, using GIS tools and properly developed software. The research, apart from the evaluation of the maximum wave run-up on the mountain side, aimed at using the numerical tool to give a theoretical interpretation of the relative importance between the different physical mechanisms which concurred in generating the catastrophe.